Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
  • H04B1/0003—Software-defined radio [SDR] systems, i.e. systems wherein components typically implemented in hardware, e.g. filters or modulators/demodulators, are implented using software, e.g. by involving an AD or DA conversion stage such that at least part of the signal processing is performed in the digital domain
  • H04B1/0007—Software-defined radio [SDR] systems, i.e. systems wherein components typically implemented in hardware, e.g. filters or modulators/demodulators, are implented using software, e.g. by involving an AD or DA conversion stage such that at least part of the signal processing is performed in the digital domain wherein the AD/DA conversion occurs at radiofrequency or intermediate frequency stage
• • G—PHYSICS
  • G11—INFORMATION STORAGE
  • G11C—STATIC STORES
  • G11C27/00—Electric analogue stores, e.g. for storing instantaneous values
  • G11C27/02—Sample-and-hold arrangements
  • G11C27/024—Sample-and-hold arrangements using a capacitive memory element
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
  • H04B1/06—Receivers
  • H04B1/16—Circuits
  • H04B1/26—Circuits for superheterodyne receivers
  • H04B1/28—Circuits for superheterodyne receivers the receiver comprising at least one semiconductor device having three or more electrodes

Definitions

• Voltage sampling is traditionally used for analog-to-digital (A/D) conversion.
• A/D analog-to-digital
• a sampling switch is placed between a signal source and a capacitor. Between two sampling moments, the capacitor voltage tracks the signal voltage accurately. At the sampling moment, the switch is turned off to hold the capacitor voltage.
• the two processes become increasingly difficult when the signal frequency increases.
• thermal noise and switching noise set a minimum allowable capacitance while the tracking speed set a maximum allowable capacitance or switch resistance. It becomes impossible when the maximum is smaller than the minimum.
• the clock jitter and finite turning-off speed (nonzero sampling aperture) make the sampling timing inaccurate.
• the bandwidth of a voltage sampling circuit must be much larger than the signal bandwidth. This makes direct sampling of high frequency radio signal extremely difficult. Sub-sampling can reduce the sampling rate but not the bandwidth of sampling circuit and not the demands on small clock jitter and small sampling aperture.
• the object of the invention is to provide an improved sampling circuit and a method of sampling an analog signal, which overcomes the above mentioned problems.
• the invention provides a charge sampling (CS) circuit, comprising a control signal generator for controlling an analog input signal to the charge sampling circuit to be integrated by an inter- grator during a sampling phase responsive to a sampling signal from the control signal generator, wherein the current of the analog input signal is integrated to an integrated charge for producing a proportional voltage or current sample at a signal output at the end of the sampling phase.
• CS charge sampling
• a more specific object of the invention is to provide a method and sampling circuit for band-pass sampling.
• a band-pass sampling (BPCS) circuit comprising a control signal generator for controlling a first and second end of a differential analog signal to be weighted by a weightmg-and-sampling (W&S) element during a W&S phase responsive to a W&S signal from the control signal generator, wherein the current of the analog signal passes through said W&S element only when said W&S signal is in a W&S phase, and said control signal generator is adapted for controlling the output signal of the W&S element to be integrated by an mtergrator during the W&S phase, wherein the current of the output signal of the W&S element is integrated to an integrated charge for producing a proportional voltage or current sample at a signal output at the end of the W&S phase.
• BPCS band-pass sampling
• Another more specific object of the invention is to provide a two-step BPCS circuit.
• a two-step BPCS circuit according to the invention which comprises a first BPCS circuit according to the invention for producing signal samples with a first sample rate; a chopping circuit for chopping the signal from the first BPCS circuit symmetrically in time at its signal output or output pair with the frequency of a clock signal equal to the first sample rate; a differential -out amplifier for amplifying the signal from the chopping circuit differentially; wherein the first signal input and the second signal input of said second BPCS are connected to the signal output pair of said amplifier (41) for producing signal samples at the signal output or output pair with a second sample rate .
• a further specific object of the invention is to provide a front-end sampling radio receiver. This is obtained by a front-end sampling radio receiver according to the invention, which comprises a low pass filter with a bandwidth up to twice the clock frequency for receiving and filtering a radio signal; a low noise amplifier for producing a differentially amplified radio signal from the filtered signal; a local oscillator for producing an
• I-clock signal at its signal output a ⁇ /2 phase shifter with a signal input connected to the local oscillator for producing a Q-clock signal at its signal output with the same amplitude and ⁇ /2 phase shift with respect to the I -clock signal; wherein two ends of the signal output pair of said low noise amplifier are respectively connected both to the first BPCS circuit and the second BPCS circuit respectively, said I -clock signal output is connected to the clock input of said first BPCS circuit, and the Q-clock signal output is connected to the clock input of the second BPCS circuit, for producing base-band I -samples of the radio signal at the signal output or output pair of the first BPCS circuit, base-band Q samples of the radio signal at the signal output or output pair of said second BPCS circuit.
• An advantage of the charge sampling circuit according to the invention is that the bandwidth of the charge sampling circuit does not have to be much larger than the signal bandwidth. Another important background is that for a radio signal, no matter how high is the carrier frequency, the signal bandwidth (the base band) remains a small fraction of the full band between DC to the carrier frequency. It is therefore unnecessary to convert the full band but just the band with the signal.
• the frequencies of the signals possibly to be sampled by the CS circuits or the BPCS circuits are higher or much higher than that of the voltage sampling circuits at a given accuracy.
• the sampling capacitors used in the CS circuits or the BPCS circuits are larger or much larger than the ones used in the voltage sampling circuits, giving advantages of low noise and low clock-and-charge feed-through.
• Each BPCS circuit is simultaneously a filter, a mixer and a sampler, which greatly simplifies a radio receiver.
• the BPCS circuits are capable of directly working at the radio frequency band, which makes a highly digitized radio receiver with front-end sampling and A/D conversion possible .
• Both the center frequency and the bandwidth of a BPCS circuit can be easily programmed. The bandwidth can be as narrow as required, equivalent to have an unlimited Q-value .
• the CS and BPCS circuits are simple and can be easily implemented in CMOS or other processes .
• This technique is very useful for the purpose of system-on-chip, which requires a simple and highly digitized architecture.
• FIG 1A is a block diagram of a first embodiment of a charge sampling (CS) circuit according to the invention
• FIG IB shows the working waveforms of the charge sampling (CS) circuit in FIG 1A
• FIG 1C shows the frequency response of the charge sampling (CS) circuit in FIG 1A
• FIG 2A is a block diagram of a first embodiment of a a band pass charge sampling (BPCS) circuit according to the invention
• FIG 2B shows the working waveforms of the band pass charge sampling (BPCS) circuit in FIG 2A
• BPCS band pass charge sampling
• FIG 3 is a block diagram of a first embodiment of a differential BPCS circuit according to the invention.
• FIG 4 is a block diagram of a first embodiment of a parallel differential BPCS circuit according to the invention
• FIG 5 shows an illustration of the filter function for the BPCS circuits according to the invention
• FIG 10 is a circuit diagram of the first embodiment of the differential BPCS circuit in FIG 3,
• FIG 11B shows the resulting frequency response of the circuit in FIG 11A
• FIG 12B is the resulting frequency response of the the circuit in FIG 12A
• FIG 13B is the resulting frequency response of the circuit in FIG 13A
• FIG 14A is a circuit diagram of a single ended active integrator
• FIG 14B is a circuit diagram of a differential active integrator
• FIG 15 is a block diagram of a two-step BPCS circuit
• FIG 16 is a block diagram of a front-end sampling radio receiver architecture.
• the present invention is a charge sampling (CS) circuit or a band-pass charge sampling (BPCS) circuit, sampling a signal by integrating its current in a given time window, and the resulting charge represents the signal sample at the center time of the window.
• CS charge sampling
• BPCS band-pass charge sampling
• FIG 1A a first embodiment of a charge sampling (CS) circuit 1 according to the invention is shown. It comprises sampling switch 2, an integrator 3 and a control signal generator 4.
• the switch 2 has a signal input, a signal output and a control input.
• An analog signal is applied to the signal input of the switch, which is the signal input of the charge sampling circuit 1, and a sampling signal is applied to the control input from the control signal generator 4.
• the swith is on, i.e the signal input is connected to the signal output of the switch, only when the sampling signal is in the sampling phase.
• the integrator 3 has a signal input, a signal output, and a control input.
• the signal output of the switch 2 is applied to the signal input of the integrator 3, and a resetting signal from the control signal generator 4 is applied to the control input of the integrator 3.
• the current of the analog input signal to the CS circuit 1 is integrated during sampling phase, and the integrated charge produces a proportional voltage or current sample at the signal output of the CS circuit at the end of the sampling phase. The sample is held until the resetting phase of the resetting signal begins, and the time interval in between is the holding phase .
• the control signal generator 4 has a clock input, which is the clock input of the CS circuit, a sampling signal output connected to the control input of the switch 2 and a resetting signal output connected to the control input of the intergrator 3 as mentioned above.
• the the integrator 3 comprises a capacitor 3-1, a resetting switch 3-2 and an optional resistor 3-3 in this embodiment.
• the integrator 3 can, however, have a different configuration in other embodiments.
• An analog signal is applied to the input of the sampling switch 2. As described, the charge sampling process involves three successive phases: resetting, sampling (ti to t ) and holding. The time from i to t 2 is defined as the sampling window.
• FIG. IB shows its working waveforms.
• the resetting phase only the resetting switch 3-2 is turned on and the capacitor 3-1 is reset.
• the sampling phase only the sampling switch 2 is turned on, and the signal current is integrated onto the capacitor 3-1.
• the time constant is large enough to be able to obtain a linear charging when the signal comes from a voltage source (the usual case) . If the on-resistance of the switch 2 is too small, the optional resistor 3-3 can be added.
• both switches are in off-state, and the output voltage of the integrator 3 is held for further use.
• a pair of interconnected CS circuits forming a differen- tial CS circuit, provide differential outputs to cancel common mode effects, using a differential input signal and sharing the control signal generator 4.
• the CS circuits or circuit pairs are used m parallel to increase the sampling rate and to make the time interval between two sampling points possibly less than the sampling window, by time- interleaving both sampling and resetting signals.
• Ii (t s ) I 1 s ⁇ n( ⁇ 1 t s + ⁇ 1 )
• the difference is (sin ((Di ⁇ t/ ⁇ j ⁇ t) ) , a sampling coefficient depending on frequency ⁇ x and ⁇ t .
• the ith frequency component has been precisely sampled at time t s . Since all frequency components are sampled at t s , the total charge on the capacitor naturally represents the signal sample at t s , i.e. t s is the equivalent sampling time point .
• the frequency response of the CS circuit depends on the function sin (cOjA /cujAt ) , shown in FIG. 1C.
• FIG. 2B shows the working waveforms.
• the integrator is reset.
• Each sampling phase includes n clock cycles forming a sampling window.
• the signal current through W&S element equals zero outside the sampling window and is weighted according to the weighting function (constant, linear, Gauss or other fuctions) within the sampling window.
• the weighting function depends on the combination of the W&S element 6 and the W&S signal. The three W&S signals shown in FIG.
• a differential BPCS circuit 8 is shown in FIG 3. It comprises four switches 2A, 2B, 2C and 2D, a differential W&S (D-W&S) element 9, a differential integrator 10, and a control signal generator 7, as connected.
• the shown type of D-W&S element 9 comprises two parallel W&S elements 6A and 6B, and the shown type of differential integrator comprises two parallel integrators 3A and 3B.
• the D-W&S element 9 and the differential integrator 10 may be in other types.
• the differential BPCS circuit 8 works in the same way as the single ended BPCS circuit 5 except to produce two outputs differentially.
• the differential BPCS circuit 8 effectively cancels the common mode effects and gives more accurate results .
• FIG 4 shows a parallel differential BPCS circuit 11.
• It comprises four switches 2A, 2B, 2C and 2D, a number of D-W&S elements 9A, 9B, ... , 9X, a number of differential integrators 10A, 10B, ... , 10X, a multiplexer (MUX) 12 and a control signal generator 13, as connected.
• Each pair of the D-W &S element and the differential integrator, 9A+10A, 9B+10B, ..., 9X+10X, together with the switches 2A, 2B, 2C and 2D work in the same way as the differential BPCS circuit 8.
• the W&S signals and the resetting signals to these pairs, generated by the control signal generator 13, are evenly time-interleaved.
• the MUX 12 multiplexes the outputs of the differential integrators 10A, 10B, ... , 10X to the differential outputs when they are in the holding phase, controlled by the multiplexing signals from the control signal generator 13.
• the parallel BPCS circuit gives a higher sampling rate and makes the time interval between two successive sampling points possibly less than the sampling window. If switches 2C and 2D are removed, and the differential W&S elements and the differential integrators are replaced by single-ended versions, it becomes a parallel single-ended BPCS circuit.
• FIG 5 A filter function of the BPCS circuits is illustrated in FIG 5. From top-down, the frequency increases from DC to 3f, c where f c is the clock frequency. Note that during the negative clock phase the same signal is connected oppo- sitely, which is reflected in the diagram by changing the signal sign.
• the normalized amplitudes of resulting charges, i.e. the sums of the areas, integrated in n clock cycles are listed in FIG 5 respectively. It is obvious that for input signals with frequencies much higher or lower than f c , the charges cancel each other almost completely, resulting in nearly zero output. For input signals with certain frequencies like f c /4, f c /2, 2f c , ..., the charges are completely cancelled no matter what are their phases.
• FIG 6A An ideal frequency response of a BPCS circuit is shown in FIG 6A, which corresponds to a mathematically accurate integration of the signal current in the sampling window.
• n 10 and constant -weighting are assumed, meaning that the weight of the current is kept constant in the 10-clock-cycle sampling window.
• the output frequency f out equals
• the same output frequency is obtained for input frequencies fini ( ⁇ (2p-l)f c ) and f ⁇ n2 (>(2p-l)f c ) when (2p-l)f c - (2p-l) f c , but their phases are different.
• the amplitudes of far-end frequency components are reduced with the increase of n, but the maximum adjacent peaks in both cases remain almost unchanged, around -13 dB.
• Linear-weighting means that during the sampling phase the weight of the current is first linearly increase and then linearly decrease, symmetric to the center of the sampling window.
• the amplitudes of far-end frequency components are rapidly reduced with the increase of n.
• the maximum adjacent peaks are reduced to -26 dB and -27 dB respectively, compared to those of the constant-weighting cases.
• FIG 9A and FIG 9B the ideal frequency responses of a Gauss-weighting BPCS circuit are shown.
• Gauss-weighting means that during the sampling phase the weight of the current varies according to the Gauss function exp(-t 2 /2 ⁇ 2 ) for a given ⁇ , symmetric to the center of the sampling window.
• the 3 dB bandwidths are both 0.025f c .
• the 3 dB bandwidths are both 0.0025f c .
• the amplitudes of far-end frequency components and the adjacent peaks are substantially reduced with the Gauss-weighting .
• the maximum adjacent peaks are in the range of -61 dB to -78 dB .
• the clocked switches are n-MOS transistors 15A, 15B,
• the W&S elements are n-MOS transistors 16A and 16B.
• the resetting switches are n-MOS transistors 18A and 18B.
• the capacitors are on-chip MOS capacitors 17A and 17B.
• the clocks are in sinuous waves but quasi-square waves can also be used.
• the implementation 14 works in all CMOS processes. Parameters of a 0.8 ⁇ m CMOS process, however, is used in the HSPICE simulations. The following three implementations are based on the implementation 14 with particular component values and W&S signal parameters.
• the clocked switches are n-MOS transistors 20A, 20B, 20C and 20D.
• the W&S elements are n-MOS transistors 21A and 21B.
• the resetting switches are n-MOS transistors 23A and 23B. They all have the minimum size, 2 ⁇ m/0.8 ⁇ m (width/length) .
• the capacitors are MOS capacitors 22A and 22B, both 40 pF.
• the maximum differential output sample voltage is around 100 V.
• the clocked switches are n-MOS transistors 25A, 25B, 25C and 25D, all having an increased size of 10 ⁇ m/0.8 ⁇ m. This makes the signal currents dominated by the W&S elements not the switches.
• the W&S elements are n-MOS transistors 21A and 21B, 2 ⁇ m/0.8 ⁇ m.
• the resetting switches are n-MOS transistors 23A and 23B, 2 ⁇ m/0.8 ⁇ m.
• the capacitors are MOS capacitors 22A and 22B, both 40 pF .
• the maximum differential output sample voltage is around 100 mV.
• the maximum adjacent peak is -30 dB, lower than that of the theoretical response. This is because the conductance of n-MOS transistors 21A or 21B does not vary linearly with the linear W&S signal.
• the actual weighting function is somewhere between linear and Gauss .
• the clocked switches are n-MOS transistors 25A, 25B, 25C and 25D, 10 ⁇ m/0.8 ⁇ m.
• the W&S elements are n-MOS transistors 27A and 27B, 2 ⁇ m/16 ⁇ m. Note that the lengths of 27A and 27B are increased to 16 ⁇ m to limit the signal current and the capacitor voltage dung such a long charging period (599 ns) .
• the resetting switches are n-MOS transistors 23A and 23B, 2 ⁇ m/0.8 ⁇ m.
• the capacitors are MOS capacitors 28A and 28B, both 20 pF.
• the simulated frequency response is basically in accordance with the theoretical frequency response. Both have MHz.
• the maximum adjacent peak of the implementation 26 is -30 dB, lower than that of the theoretical response.
• FIG 14A and FIG 14B show active integrators for improving output swing and linearity, respectively.
• a single ended active integrator 29 is shown in FIG 14A.
• the inverter 35 produces an inverted resetting signal with a delay, using the resetting signal as the input, to control the switch 33 while the resetting signal controls the switches 32 and 34. During the resetting phase, the switches 32 and 34 are turned on, and the switch 33 is turned off. The voltage of capacitor 31 is reset to the input offset voltage of the amplifier 30.
• a differential active integrator 36 is shown in FIG 14B. It comprises a differential-in-differential-out amplifier 37, two capacitors 31A and 31B, an inverter 35, and switches 32A, 32B, 33A, 33B, 34A and 34B. It works basically in the same way as the integrator 29 except uses a differential input signal and gives differential outputs.
• the integrator 29 can replace the integrator 3 in FIG 1A while the integrator 36 can replace the integrator 10 in FIG 3.
• FIG 15 shows a two-step BPCS circuit 38. It comprises a first BPCS circuit 39, a chopping circuit 40, an amplifier 41, a second BPCS circuit 42, and a clock signal generator 43 generating a second clock.
• the first BPCS circuit 39 and the second BPCS circuit 42 can be any type of the BPCS circuits 5, 8, 11, 19, 24 and 26.
• To the first BPCS circuit 39 two ends of a differential analog signal are applied to its two inputs respectively, and a first clock is applied to its clock input. Signal samples with a first sample rate are produced from the first BPCS circuit 39 and fed to the chopping circuit 40. The samples are chopped symmetrically m time, controlled by the second clock.
• the chopped signal with a new carrier frequency equal to the chopping frequency is fed to the amplifier 41, and the amplified differential signals are fed to two inputs of second BPCS circuit 42 respectively. Controlled by the second clock, the second BPCS circuit 42 produces the final sample output with a second sample rate.
• the two-step BPCS circuit 38 gives flexibility in performance trade-off. BPCS circuits in more steps can be built based on the two-step BPCS circuit 38.
• a front -end sampling radio receiver architecture 44 is shown in FIG 16. It comprises a low pass filter 45 with f Pass ⁇ 2f c a differential -out low noise amplifier (LNA) 46, two BPCS circuits 47A and 47B, a 90° phase shifter 48, and a local oscillator 49.
• the radio signal from antenna is applied to the input of the low pass filter 45.
• the frequency components above 2f c are greatly attenuated.
• the output of the low pass filter 45 is fed to the LNA 46 to produce differential outputs with a large enough amplitude.
• the differential outputs are fed to the inputs of BPCS circuits 47A and 47B simultaneously.
• the I-clock signal produced by the local oscillator 49 is fed to the BPCS circuit 47A while the Q-clock signal reproduced by the 90° phase shifter 48 from the I-clock signal is fed to the BPCS circuit 47B.
• the BPCS circuits 47A and 47B produce I-samples and Q-samples respectively.
• the sample outputs can be either converted to digital data immediately or further treated.
• the BPCS circuits 47A and 47B can be any of the BPCS circuits 5, 8, 11, 19, 24 and 26.
• the mte- grators m these circuits can be either passive integrators or active integrators.
• the radio receiver architecture 44 has filtering, mixing and sampling functions simultaneously at the front-end, which relaxes the performance demands on A/D conversion, avoids analog filters, and highly utilizes the capability of DSP. In principle, any narrow bandwidth, i.e. any high Q value, is possible. The center frequency of the filtering function can be easily programmed. It is indeed a superior radio receiver architecture with a wide application scope.
• the sampling capacitors used in the CS and the BPCS circuits are much larger than that used in a voltage sampling circuit, resulting in low noise and low charge and clock feed-through.
• the BPCS circuit is simultaneously a filter, a mixer and a sampler, capable of working at radio frequencies.
• the center frequency, the bandwidth and the adjacent selectivity can be set by the clock frequency, the number n and the shape of W&S signal, particularly useful for front-end sampling radio receiver and system-on-chip.

Landscapes

• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• Analogue/Digital Conversion (AREA)
• Networks Using Active Elements (AREA)
• Transition And Organic Metals Composition Catalysts For Addition Polymerization (AREA)
• Glass Compositions (AREA)
• Peptides Or Proteins (AREA)
• Separation By Low-Temperature Treatments (AREA)

Priority Applications (5)

Applications Claiming Priority (2)

Publications (1)

Family

ID=20417207

Family Applications (1)

Country Status (9)

Cited By (4)

Families Citing this family (13)

Citations (7)

Family Cites Families (23)

• 1999
  • 1999-09-28 SE SE9903532A patent/SE9903532D0/xx unknown
• 2000
  • 2000-09-25 WO PCT/SE2000/001854 patent/WO2001024192A1/en active Application Filing
  • 2000-09-25 AT AT00968260T patent/ATE506677T1/de not_active IP Right Cessation
  • 2000-09-25 DE DE60045867T patent/DE60045867D1/de not_active Expired - Lifetime
  • 2000-09-25 AU AU78201/00A patent/AU7820100A/en not_active Abandoned
  • 2000-09-25 EP EP00968260A patent/EP1221166B1/en not_active Expired - Lifetime
  • 2000-09-25 JP JP2001527292A patent/JP4685310B2/ja not_active Expired - Fee Related
  • 2000-09-25 EP EP10168098A patent/EP2228800A3/en not_active Withdrawn
  • 2000-09-25 CN CNA2004100598399A patent/CN1551505A/zh active Pending
  • 2000-09-25 CN CNB008135592A patent/CN1174431C/zh not_active Expired - Fee Related
  • 2000-09-28 US US09/672,803 patent/US7053673B1/en not_active Expired - Fee Related
• 2005
  • 2005-04-06 US US11/099,488 patent/US7023245B2/en not_active Expired - Fee Related
  • 2005-04-06 US US11/099,460 patent/US8035421B2/en not_active Expired - Fee Related
• 2010
  • 2010-12-09 JP JP2010274877A patent/JP4875201B2/ja not_active Expired - Fee Related

Patent Citations (7)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events